Comprehensive nuclear materials 3 02 nitride fuel

Comprehensive nuclear materials 3 02 nitride fuel Comprehensive nuclear materials 3 02 nitride fuel Comprehensive nuclear materials 3 02 nitride fuel Comprehensive nuclear materials 3 02 nitride fuel Comprehensive nuclear materials 3 02 nitride fuel Comprehensive nuclear materials 3 02 nitride fuel Comprehensive nuclear materials 3 02 nitride fuel

Y Arai

Nuclear Science and Engineering Directorate, Japan Atomic Energy Agency, Ibaraki, Japan

ß 2012 Elsevier Ltd All rights reserved.

Abbreviations

ADS Accelerator-driven system

CAPRA Fast reactor operated to burn rather than

breed plutonium

EPMA Electron probe microanalyzer

FCCI Fuel–clad chemical interaction

FCMI Fuel–clad mechanical interaction

FIMA Fission per initial metal atom

HLLW High-level radioactive liquid waste

HLW High-level radioactive waste

ITU European Institute for Transuranium

Elements

JAEA Japan Atomic Energy Agency

LINEX Direct synthesis of actinide nitrides in the

salt by the reaction with Li3N

LOF Loss of flow

MOX Uranium and plutonium mixed oxide

PIE Postirradiation examination

PSI Paul-Scherrer Institute

PUREX Plutonium uranium recovery by extraction

SPS Spark-plasma sintering

TD Theoretical density TOP Transient overpower XRD X-ray diffraction

3.02.1 Introduction Nitride fuel has been proposed as an advanced fuel for fast reactors and developed since the 1960s in almost the whole nuclear world In this case, nitride fuel stands for a solid solution of uranium mononi-tride (UN) and plutonium mononimononi-tride (PuN), namely (U,Pu)N, in which the Pu/(Uþ Pu) molar ratio roughly ranges from 0.15 to 0.25 In addition,

UN was developed as a potential fuel for space reac-tors in the United States Although the interest in nitride fuel subsided under a global circumstance

of slowdown of fast reactor programs in the 1980s, the solid solution of UN, PuN, and minor actinide (MA; Np, Am, and Cm) mononitride, (U,Pu,MA)N, has been proposed as one of the candidate fuels for Gen IV-type fast reactors Furthermore, as a

41

dedicated fuel for MA transmutation systems such as

an accelerator-driven system (ADS), U-free nitride

fuel, such as (Pu,MA)N diluted by ZrN, has been

studied mainly in Japan

At the beginning of the nuclear era, the

develop-ment of fast reactor fuel cycles was centered on

breeding ratio and doubling time The reason was

that metallic fuel, the binary or certain ternary alloy

of U and Pu, was adopted in the first generation of

fast reactors The metallic fuel, however, had

disad-vantages for commercial use, such as anisotropic

crystal structure, low melting temperature with

phase transformations, and high fission product (FP)

gas-induced swelling So a solid solution of uranium

dioxide (UO2) and plutonium dioxide (PuO2),

namely (U,Pu)O2(MOX), has been a reference fuel

for fast reactors and used in many test and prototype

reactors all over the world, although the breeding

ratio is smaller and the doubling time is longer than

those of metallic fuel

On the other hand, nitride fuel, as well as carbide

fuel, has the advantages of both metallic fuel and

oxide fuel as shown inTable 1 It has a high thermal

conductivity and high metal atom density like

metal-lic fuel, while it has a high melting temperature and

isotropic crystal structure like oxide fuel These

char-acteristics led to the motivation for developing

nitride fuel for fast reactors because the high thermal

conductivity and high melting temperature allow a

high linear power operation; alternatively, the

large-diameter fuel pins can be used for a given linear

power The high metal atom density allows a low

fissile material inventory with flexible core design

and good neutron economics, leading to an improved

breeding ratio and doubling time

However, the development of nitride fuel has

fallen behind that of carbide fuel, which has similar

physical and chemical properties The reason

in-cludes an unexploited fuel fabrication process and

the high neutron capture cross section of 14N (99.6% abundance in natural nitrogen) deteriorating neutron economics However, the fuel fabrication process has improved since the late 1980s, and the breeding ratio and doubling time have not been the center of the development of fast reactor fuel cycles Furthermore, it was found that nitride fuel is less hygroscopic in nature than carbide fuel, which will be advantageous for technological development

It was also found that nitride fuel dissolves well in nitric acid without any formation of Pu oxalate, which will be compatible with hydrochemical repro-cessing technology such as the PUREX process So since the late 1980s, the global interest has moved from carbide fuel to nitride fuel

We can find two distinguished monographs about nitride fuel: one is written by Matzke1 published in

1986 and the other by Blank2in 1994 These mono-graphs describe nitride fuel and carbide fuel as MX-type fuel (X¼ N or C) for fast reactors in detail from scientific and technological viewpoints It should also

be mentioned that (U,Pu)N fuel with high Pu con-tent, in which the Pu/(Uþ Pu) molar ratio is roughly 0.45–0.55, was studied in France as a fast reactor fuel for incineration of Pu in the 1990s The good disso-lution in nitric acid and stable crystal structure even

at high Pu content led to the potential CAPRA core with (U,Pu)N fuel for incineration of Pu.3Although not being described in this chapter, another interest-ing aspect of nitride fuel was pointed by Lyonet al.,4 who indicated the superior safety margin in case of hypothetical loss of flow (LOF) and transient over-power (TOP) events

In a space reactor program called SP-100 in the United States, UN with highly enriched 235U was chosen as a reference fuel because it has the most favorable properties and will show the best perfor-mance for space reactor fuels.5 An extensive work was carried out in SP-100 program and Hayes et al

Table 1 Comparison of typical properties between oxide, metallic, and nitride fuels for fast reactors

Chemical composition (U 0.8 Pu 0.2 )O 2 U–19Pu–10Zr (wt.%) (U 0.8 Pu 0.2 )N

Thermal conductivity (W m1K1)

a At 0.1 MPa N pressure.

summarized the physical,6 mechanical,7 transport,8

and thermodynamic properties9of UN, while FP gas

release and swelling of UN were summarized by

Storms10 and Ross et al.,11

respectively Ross et al.12 also compiled and analyzed the thermal conductivity

data of UN On the other hand, the diffusional and

mechanical properties were reviewed by Routbort

et al.13

previously

Since the late 1990s, the partitioning and

transmu-tation of MA has attracted global interest It may

contribute to the decrease of toxicity of high-level

radioactive waste (HLW) and the mitigation of burden

for its final disposal Several transmutation systems

and MA-containing fuels have been proposed so far

Among them, the Japan Atomic Energy Agency

( JAEA) proposed a subcritical ADS as a

transmuta-tion system and MA nitride fuel as a dedicated fuel for

transmutation.14 Besides the thermal and neutronic

properties, the mutual solubility of actinide

mononi-trides in a wide range of composition and combination

becomes an advantage of the fuel with high MA

con-tent Fabrication of MA nitride fuel and its property

measurements have been carried out in JAEA.15–18

In this chapter, fabrication of nitride fuel and its

irradiation behavior are summarized in Sections

3.02.2 and 3.02.3, respectively A brief description

about reprocessing of spent nitride fuel is given in

Section 3.02.4, because the reprocessing technologies

are closely related with the specific issues of nitride fuel

as14C formation from natural nitrogen and15N

enrich-ment On the other hand, properties of nitride fuel are

described in Chapter 2.03, Thermodynamic and

Thermophysical Properties of the Actinide Nitrides

In addition, an outlook of nitride fuel is briefly given

inSection 3.02.5

3.02.2 Fabrication of Nitride Fuel

3.02.2.1 Actinide Nitride Compounds

Although nitride fuel usually stands for a

mononi-tride or its solid solution, such as UN and (U,Pu)N,

higher nitrides other than mononitrides exist in the

Th–N and U–N binary systems.Table 2summarizes

the crystal structures and lattice parameters of

acti-nide nitrides reported in the Th–N, U–N, Np–N,

Pu–N, Am–N, and Cm–N binary systems The

binary U–N and Pu–N, and ternary U–Pu–N

sys-tems were investigated and reviewed by Holleck,19,20

Tagawa,21and Potter22in detail

The ternary U–Pu–N phase diagram at 1273 K

in Matzke’s monograph,1 originally calculated by

Holleck,19is shown inFigure 1 The system is char-acterized by a complete solubility of UN and PuN

It is considered that (U,Pu)N phase has a narrow composition range of the N/(Uþ Pu) molar ratio Although Pu2N3does not exist in the Pu–N system,

a sesquinitride phase was identified in the U–Pu–N system at a Pu/(Uþ Pu) molar ratio of 0.15.23

As seen

in Table 2, actinide mononitrides have the same crystal structure with similar lattice parameters ex-cept for ThN, which leads to the mutual solubility

In a mononitride lattice with NaCl-type structure, small nitrogen atoms are incorporated into a dense face-centered cubic packing of metal atoms

Table 2 Crystal structures and lattice parameters of nitrides of Th, U, Np, Pu, Am, and Cm

Compounds Structure Lattice parameter

(nm)

Th 3 N 4 Th 3 P 4 -type

hexagonal

a ¼ 0.3871

c ¼ 2.7385

a-U 2 N3þ x Mn 2 O 3 -type bcc 1.0685 b-U 2 N3 x La 2 O 3 -type

hexagonal

a ¼ 0.3696

c ¼ 0.5840

UN2 x CaF 2 -type fcc 0.531

N

1 atm

U – Pu – N

UN

β + (U,Pu)N

α + β + (U,Pu)N

+ (U,Pu)N PuN

Figure 1 Ternary U–Pu–N phase diagram at 1273 K Reproduced from Matzke, Hj Science of Advanced LMFBR Fuels; North-Holland: Amsterdam, 1986.

3.02.2.2 Preparation from Metal or Hydride

Nitride preparation methods from metal or hydride

were investigated mainly in the 1960s They include

the nitridation of U or Pu metal in N2 or NH3 at

about 1073–1173 K, arc-melting of U or Pu metal

under N2pressure, nitridation of fine grained U or

Pu powder formed by the decomposition of hydrides

with N2or NH3and direct reaction of UH3or PuH2.7

with N2or NH3 In the case of uranium nitrides, the

products were often U2N3, which was subsequently

decomposed to UN and N2

These reactions are exothermic and should be

carried out slowly by temperature cycling for better

control of the products Furthermore, these methods

necessitate a high-purity inert gas atmosphere, since

the fine-grained powders of metal, hydride, and

nitride are chemically active and likely to react with

moisture and oxygen in air even at room temperature

So it is difficult to apply the metal or hydride route to

a technological fuel production line and these

meth-ods were restricted to a laboratory scale experiment

3.02.2.3 Carbothermic Reduction

Carbothermic reduction is the most widely used

pro-cess for preparing nitride fuel The starting material

is a dioxide and carbon, and the general reaction is

expressed as

MO2þ 2C þ 0:5N2¼ MN þ 2CO ½I

where M represents an actinide element, such as

U and Pu The mixture of dioxide and carbon is

heated in N2 gas stream, usually at 1773–1973 K It

is considered that the carbothermic reduction could

be applied in a technological production line as well

as in a laboratory scale experiment, in contrast to the

metal or hydride route.24Furthermore, homogeneous

products can be obtained by carbothermic reduction

However, high amounts of oxygen, up to several

thousand parts per million, are likely to remain in the

products as impurity in case the initial carbon to

diox-ide mixing molar ratio, C/MO2, is 2.0 Therefore, an

excess amount of carbon is usually added to the mixture

to reduce the oxygen content and the residual carbon is

removed from the products by heating in N2–H2

stream as CH4(25)or HCN26after carbothermic

reduc-tion The initial C/MO2 mixing ratio was chosen at

2.2–2.5 for the preparation of UN and (U,Pu)N Besides

the two-step reaction constituted by the carbothermic

reduction in N2stream and the following

decarburiza-tion in N–H stream, a one-step reaction in N–H or

NH3stream can be applied although a higher initial C/MO2mixing ratio is necessary than that for the two-step reaction For the preparation of UN and (U,Pu)N, the atmosphere is changed to Ar or He from N2 or

N2–H2 at a temperature lower than about 1673 K to prevent the formation of higher nitrides

In the case of preparation of solid solution such

as (U,Pu)N, both the reduction of the mixture of respective dioxides and the solid solution formation

of respective mononitrides can be applied Figure 2

shows the X-ray diffraction (XRD) pattern of (Np,Pu, Am,Cm)N prepared by the carbothermic reduction of the mixture of respective dioxides, from which the formation of quaternary mononitride solid solution was confirmed

Mechanism and kinetics of carbothermic reduc-tion were investigated by several authors, such as Muromuraet al.,27–30

Lindemer,31 Greenhalgh32and Bardelleet al.,26

mainly by chemical and XRD ana-lyses, and weight change measurement for UN, PuN, and (U,Pu)N Muromuraet al investigated the mech-anism of carbothermic reduction at 1693–2023 K for

UN in detail According to their results, the reaction

is divided into four stages: (1) formation of UN1  xCx from UO2, (2) decarburization of UN1  xCx, (3) for-mation of UN1  xCxwith equilibrium composition, and (4) pure UN formation They also claimed that the carbothermic reduction followed the first-order rate reaction expressed as

 lnð1  aÞ ¼ kt ½1 where a represents the reaction ratio, k the rate constant, and t the time, with an activation energy

CmN AmN PuN NpN CmN AmN PuN NpN

66 65 64 63 62 61

2q (deg)

Figure 2 X-ray diffraction pattern of (Np,Pu,Am,Cm)N prepared by carbothermic reduction 81 Reprinted with permission from OECD/NEA (2007), Actinide and Fission Product Partitioning and Transmutation, Ninth Information Exchange Meeting, Nıˆmes, France, Sept 25–29, 2006,

p 119, www.nea.fr.

of 347 kJ mol1 This value is consistent with that

reported by Greenhalgh,32 360 kJ mol1 On the

other hand, Muromuraet al claimed that the

decar-burization in N2–H2or NH3stream after

carbother-mic reduction followed the phase boundary-type rate

reaction expressed as

1 ð1  aÞ1=3¼ kt ½2

with activation energies of 285 kJ mol1 in 25%

N2–75% H2 stream and 175–185 kJ mol1 in NH3

stream, respectively

Kinetics was also investigated by

thermogravime-try for (U,Pu)N33and (U,Np)N.34The results almost

agreed with that for UN by Muromura et al.; the

carbothermic reduction in N2 stream followed

the first-order rate reaction with activation energies

of 307 kJ mol1 for (U,Pu)N and 344–385 kJ mol1

for (U,Np)N Furthermore, the decarburization for

(U,Np)N in 92% N2–8% H2 stream followed the

phase boundary-type rate equation with an apparent

activation energy of 210 kJ mol1 However, it should

be pointed out that the decarburization includes

both the removal of free carbon resulting in a

decrease in weight and the replacement of carbon

by nitrogen in carbonitride resulting in an increase

in weight

Typical impurities in nitride fuel prepared by

carbothermic reduction are oxygen and carbon

It was found that the level of impurities could be

kept lower than 1000–2000 ppm for both oxygen and

carbon by adjusting the initial C/MO2 mixing ratio

Carbonitrides such as UN1  xCxand PuN1  xCxare

characterized by complete solubility of the UN–UC

and PuN–PuC systems, while solubility limits of

hypo-thetical UO in UN and PuO in PuN were reported at

7% and 14%, respectively.35 It was reported that the

carbon impurity content in mononitride prepared by

carbothermic reduction is related to the

thermody-namic equilibrium composition of carbonitride with

free carbon under nitrogen atmosphere.17 When the

same condition of carbothermic reduction was applied

for UN, NpN, and PuN, the carbon impurity content

decreased with the increase of atomic number of

actinides Indeed, a rather high initial C/MO2

mix-ing ratio was chosen for the preparation of AmN and

(Pu,Cm)N,36,37 since the monocarbides of Am or

Cm are thermodynamically unstable

It is well known that Am-bearing species have high

vapor pressures in comparison with the other

acti-nides Vaporization of Am during fuel fabrication

pro-cess should be kept as low as possible In the case of

preparation of Am-bearing nitrides by the two-step reaction, the carbothermic reduction in N2 stream was carried out at 1573 K, which was lower than the cases for UN and (U,Pu)N by about 200 K.36,38Then the temperature was raised to 1773 K for the decarbu-rization in N2–H2 stream It is considered that the intermediate product of AmCO is likely to vaporize congruently during the carbothermic reduction On the other hand, the vaporization of Pu during car-bothermic reduction can be neglected, which is differ-ent from the preparation of Pu-bearing carbides by carbothermic reduction carried out in vacuum

3.02.2.4 Other Nitride Formation Processes

Four processes were reported for the preparation of nitride with regard to pyrochemical reprocessing

of spent fuel The first one is the direct dissolution

of spent nitride fuel in liquid Sn, followed by the pressurization with N2 It was reported that UN powder with high density sank to the bottom and could be mechanically separated from the liquid phase.39 The second and third processes concern the nitridation of actinides recovered in liquid Cd cathode by molten salt electrolysis The second one is the nitridation by N2gas bubbling, in which N2gas is passed into liquid Cd phase at 773–823 K Kasaiet al reported that they succeeded in preparing UN or

U2N3granules by the N2gas bubbling method.40 It was found, however, that the method was not applica-ble to the nitridation of Pu in liquid Cd because of the thermodynamic stabilization of Pu in liquid Cd phase.41 On the other hand, the third one is the nitridation–distillation combined reaction, in which the liquid Cd cathode-containing actinides are heated in N2 stream at 973 K In this method, the nitridation of actinides and distillation of Cd proceed simultaneously Preparation of (U,Pu)N, PuN, and AmN has been reported so far by the nitridation– distillation combined method.41,42The fourth one is called LINEX process, in which actinides dissolved

in the chloride molten salt are converted to nitride by the direct reaction with Li3N.43

In addition, a new process was reported by Yeamans

et al.44 They successfully synthesized UN from UO2by making it react first with NH3(HF)2at ambient tem-perature to form (NH4)4UF8, and then with NH3 at

1073 K to UN2, followed by the decomposition to UN

at 1373 K in Ar This method has the advantage of a low-temperature operation in comparison with the carbothermic reduction of dioxides

3.02.2.5 Nitride Pellet Fabrication

Nitride fuel pellets are usually prepared by a classical

powder metallurgical manner; the product of

car-bothermic reduction is ground to powder by use of

a ball mill, pressed into green pellets and sintered in a

furnace at 1923–2023 K An organic binder is

some-times added to the ground powder to facilitate the

pressing Finally, the diameter of sintered pellets is

adjusted by use of a centerless grinder As is

men-tioned later, one of the characteristics of nitride fuel is

that both He- and Na-bonded pins can be applied In

general, an He-bonded fuel pin is characterized by

low-density pellets (i.e., 80–85% of theoretical

den-sity (TD)) and a small gap width between pellets and

cladding tube, whereas a Na-bonded fuel pin is

char-acterized by high-density pellets (i.e., >90% TD)

and a large gap width

Actinide nitride powder has a low sinter-ability in

comparison with that of oxide or carbide powder,

which is derived from a low diffusion rate of metal

atoms in mononitrides So a rather high sintering

temperature (i.e., T >1973 K) is necessary for

pre-paring dense UN or (U,Pu)N pellets higher than 90%

TD.45 Although a small amount of Ni powder is an

effective sintering aid for carbide fuel, it is not

appli-cable to nitride fuel On the other hand, Bernardet al.,

reported that oxygen impurities tend to promote the

sintering of (U,Pu)N pellets.24However, the increase

of oxygen impurities in UN and (U,Pu)N up to

1 wt% resulted in the decrease of density and grain

size of sintered pellets.46Microstructures of (U,Pu)N

pellets with different oxygen impurity contents are

shown inFigure 3

Sintering atmosphere also affects the sintered

density of nitride fuel pellets It was reported that

sintering in high N2partial pressure, such as in N2or

N2–H2stream, resulted in lower density than sinter-ing in low N2partial pressure, such as in Ar or Ar–H2 stream.45,47 This is an opposite tendency of the self-diffusion coefficient of Pu in (U,Pu)N at differ-ent N2 partial pressures.48 The residual oxygen impurity contents might affect the density of pellets sintered in different atmospheres On the other hand, sintering in N2 or N2–H2 stream is indispensable for Am-bearing nitride pellets from the viewpoint

of mitigating loss of Am by evaporation It was reported that the density higher than 85% TD was attained for (Np,Am)N and (Pu,Am)N pellets by sintering in N2–H2 stream at temperatures lower than 1953 K.49

In addition to the classical powder metallurgical manner, a direct pressing (DP) method was proposed

by Richter et al.50

In this method, the nitride com-pacts after carbothermic reduction were not ground

to powder but directly pressed into green pellets, followed by sintering under the conventional manner The DP method has the advantage of avoiding dust production and shortening preparation period The (U,Pu)N pellets prepared by the DP method had a density of about 83% TD with levels of oxygen and carbon impurities lower than 0.1 wt%.51 The open porosity predominated in the pellets prepared by the

DP method

An isostatic hot-pressing technique was applied to fabrication of dense UN specimens for thermal and mechanical property measurements Speidel et al prepared UN pellets higher than 95% TD by conso-lidating the powder sealed in a refractory metal con-tainer under a pressure of 6.9 MPa at 1753–1813 K.52 Furthermore, a spark-plasma sintering (SPS) method for nitride fuel has been applied to preparation in a laboratory scale experiment recently.53 The SPS

(U,Pu)N pellet containing 0.21 wt% oxygen

(U,Pu)N pellet containing 0.99 wt% oxygen

20μm Oxide

Figure 3 Microstructures of (U,Pu)N pellets with different oxygen impurity contents Reproduced from Arai, Y.; Morihira, M.; Ohmichi, T J Nucl Mater 1993, 202, 70–78.

method is a kind of pressure-assisted sintering that

utilizes an electric current The method has the

advan-tage of obtaining dense pellets at a drastically lower

sintering temperature and a shorter sintering time

than those of the conventional methods

3.02.2.6 Nitride Particle Fabrication

Nitride particle fabrication method was vigorously

developed in the Paul-Scherrer Institute (PSI) of

Switzerland,54,55then followed by India56and Japan.57

The starting material is usually a nitric solution of

actinides and this method has the advantage of avoiding

dust production and feasibility of remote operation in

comparison with the conventional powder process

The nitride particles prepared may be directly filled

into fuel pin (sphere-pac fuel) or pressed and sintered

to fuel pellets

The production of microspheres is carried out by

a so-called sol–gel process The feed solution is

mixed with an aqueous solution of gelation agent,

urea, dispersed carbon black, and surfactants

Differ-ent size of microspheres can be obtained by changing

the nozzle used for microspheres production Besides

the external gelation process using gelation agent, the

internal gelation process developed by PSI consists of

falling the droplets of feed material into hot silicon

oil for microspheres production After washing,

dry-ing and calcindry-ing to MO2þ C microspheres, they are

subjected to carbothermic reduction In the case of

preparing sphere-pac fuels, the carbothermic

reduc-tion is carried out at higher temperature than the

conventional powder process to obtain dense nitride

particles by reaction sintering

The sol–gel process is proposed for the

prepara-tion of nitride fuel for the transmutaprepara-tion of MA under

the double-strata fuel cycle concept.14 In this

con-cept, MA partitioned from high-level liquid waste

(HLLW) in a reprocessing plant is converted to

nitride microspheres by the sol–gel process and

car-bothermic reduction, followed by mixing with

dilu-ent materials and sintering for pellet preparation

3.02.3 Irradiation Behavior of Nitride Fuel

3.02.3.1 Irradiation Experience The irradiation experience of nitride fuel is rather limited in comparison with the other fuels for fast reactors, such as oxide, metallic, and carbide fuels Especially, the number of (U,Pu)N fuel pins irradiated

in fast reactors so far is smaller than 200 all over the world, which is summarized inTable 3 The highest burnup was attained in the irradiation test in the EBR-II fast reactor, but still lower than 10% of fission per initial metal atom (FIMA).58On the other hand, high burnups, that is,>15% FIMA, were attained in thermal reactors, such as ETR in the United States62 and HFR in the Netherlands.63 Most of them were irradiated in instrumented capsules

In the United States, following the capsule irradi-ation in ETR and EBR-II, 3 subassemblies consti-tuted by 57 (U,Pu)N fuel pins were irradiated in EBR-II,64whereas in Europe, more than 10 (U,Pu)N fuel pins were irradiated in fast test reactors, such

as DFR, RAPSODIE, and PHENIX.59,60Besides, in Japan, two (U,Pu)N fuel pins were irradiated in fast test reactor JOYO.61

With regard to nitride fuel other than (U,Pu)N, five subassemblies of 235U-enriched UN fuel were irradiated to about 9% FIMA in BR-10 in the 1980s.65 In addition, nitride fuels for the transmuta-tion of MA have been subjected to the irradiatransmuta-tion tests recently Besides (U,Pu,Np,Am)N and (Pu,Am, Zr)N fuels irradiated in PHENIX,66 (Pu,Zr)N fuels were irradiated in Russia67and Japan.68

3.02.3.2 Fuel Design There are two typical bonding concepts of (U,Pu)N fuel pins for fast reactors: one is Na bonding and the other is He bonding Since (U,Pu)N fuel is compati-ble with liquid Na at operating temperatures, the gap between fuel pellets and cladding tube can be filled with liquid Na as well as gaseous He In a sense of

Table 3 Irradiation tests of (U,Pu)N fuel carried out in fast reactors

Reactor Bonding Max linear power (kW m1) Max burnup (% FIMA) References

liquid metal, liquid Li bonding was also suggested

for UN-fueled space reactors In a He-bonding

con-cept, the gap is filled with He of atmospheric

pres-sure Besides the pellet-type fuel, vibropac (U,Pu)N

fuel pins were irradiated in DFR by use of He for

bonding gas.59

A Na-bonding concept is characterized by a large

gap width (i.e., >0.5 mm) between fuel pellets and

cladding tube and a high density of fuel pellets (i.e.,

>90% TD) This concept has the advantage of

keeping the fuel temperature relatively low due to

good thermal conductivity of liquid Na

Further-more, the temperature of fuel pellets is considered

as quasiconstant A shroud tube was sometimes used

in order to maintain the fuel fragments in their

origi-nal geometry On the other hand, the disadvantage

of a Na-bonding concept includes the difficulty in

fuel pin fabrication and spent fuel reprocessing

Fur-thermore, with regard to safety consideration, the

possibility of loss of Na in a breached pin has to be

evaluated

At present, a He-bonding concept is considered

as the reference for (U,Pu)N fuel A He-bonding

concept is characterized by a small gap width (i.e.,

<0.2 mm) and a low density of fuel pellets (i.e.,

80–85% TD) The temperature of fuel pellets

becomes high in comparison with the fuel with Na

bonding, especially at an early stage of irradiation

However, the small gap is closed by free swelling of

fuel pellets at a burnup of 2–3% FIMA, which

enhances the gap conductance and lowers the fuel

temperature A schematic change in fuel

tempera-ture for He-bonded (U,Pu)C fuel pin is illustrated

in Figure 4, which is also applicable to (U,Pu)N

fuel pin.69 The irradiation period A, as shown in

Figure 4, corresponds to the first rise of power and

lasts for one to several days, the period B the

resin-tering of pellets center and closure of He gap, and

the period C the quasistate irradiation period in

which the fuel–clad mechanical interaction (FCMI)

starts In order to accommodate the swelling and

mitigate the strain on the cladding tube at burnup

progressing, a rather low smear density (i.e., 75–80%

TD) is adopted for He-bonded fuel pin

Blank proposed the ‘cold fuel concept’ for

MX-type fuel, in which the maximum fuel temperature is

kept lower than one-half or one-third of the melting

temperature in Kelvin.69 If this concept is realized,

both low fission gas release and mild restructuring

and mild swelling characteristics can be compatible

in both Na-bonded and He-bonded fuels

3.02.3.3 Chemical Forms of FP Chemical forms of FP in nitride fuel were evaluated

by a thermodynamic equilibrium calculation and burnup-simulated experiments70,71as well as postir-radiation examinations (PIE).72These results agreed with each other in general but it is difficult to identify the phases other than mononitride by XRD or metal-lographic analysis even at a burnup higher than 10% FIMA

Table 4shows the most probable chemical forms

of FP in the irradiated (U,Pu)N fuel Among them,

Time scale extented 400

800 1200 1600

6 5 4 3 2 1 0

Burnup (at.%) Figure 4 Schematic change in temperature for MX-type fuel pin Reproduced from Blank, C J Less Common Met.

1986, 121, 583–603.

Table 4 Chemical forms of typical FP in the irradiated (U,Pu)N fuel

Element a Chemical

forms

Element Chemical

forms

Pd (U,Pu)(Pd,Ru,Rh) 3 Pm PmN

Rh (U,Pu)(Pd,Ru,Rh) 3 Ru (U,Pu)(Pd,Ru,Rh) 3

a Elements with concentrations <0.08 at.% at burnup of 10% FIMA are not shown.

gaseous FP such as Xe and Kr exist as an elementary

state Semivolatile FP such as Cs, I, and Te are likely

to exist as an elementary state or compounds such as

CsI and CsTe Rare earth elements such as Nd, Ce,

Pr, and Y, and Zr and Nb are considered to be

dissolved in (U,Pu)N and form a mononitride solid

solution On the other hand, Mo and Tc are

consid-ered to exist in an elementary state together, and

Ba and Sr are considered to form lower nitrides

such as Ba3N2and Sr3N2 Platinum group elements

such as Pd, Ru, and Rh are likely to form an

interme-tallic compound, (U,Pu)(Pd,Ru,Rh)3, in the irradiated

nitride fuel

The change of N/(U + Pu) ratio in mononitride

phase was evaluated and Bradbury et al.71

reported that it increased by 2.1% at a burnup of 10% FIMA

Furthermore, it should be mentioned that the

chem-ical forms of FP are also influenced by the oxygen

and carbon impurity contents, that is, rare earth

ele-ments are likely to form oxide precipitates and Zr is

likely to form ZrC dissolved in mononitride phase

It was reported that the lattice parameter of the

mononitride phase of (U,Pu)N fuel did not

signifi-cantly change with burnup progressing.70 This

ten-dency was explained by the compensation of the

increase in lattice parameter due to the dissolution

of rare earth elements and the decrease due to the

dissolution of ZrN

3.02.3.4 Restructuring

Because of relatively low fuel temperature and

tem-perature gradient, the restructuring of (U,Pu)N fuel

is mild in comparison with MOX fuel for fast

reac-tors However, in the He-bonded (U,Pu)N fuel

irra-diated at high linear power, a distinct restructuring

was observed, in which three structural zones shown

inFigure 5were identified by Matzke.1Zone I found

in the central of the fuel pellet was characterized by very porous structure The pores were grown to roughly the grain size and FP gas release was high

A small central hole was sometimes observed in Zone I of the He-bonded (U,Pu)N fuel.59,60 The mechanism of formation is, however, different from that in MOX fuel; according to Coquerelleet al.72

it results from the migration of lenticular pores up the radial temperature gradient in MOX fuel, whereas it

is apparently created by an in-pile resintering mecha-nism in (U,Pu)N fuel Zone II was characterized by pseudocolumnar grains observed in MOX fuel How-ever, it was not observed in (U,Pu)N but only in (U,Pu)C irradiated at a linear power higher than

100 kW m1 Zone III was characterized by the struc-ture accompanied with grain growth, grain boundary bubbles, and healing of cracks FP gas release was relatively high and swelling of (U,Pu)N fuel was mostly responsible in this zone On the other hand, Zone IV had the as-fabricated structure In the case of low-density pellets, slight densification occurred because of in-pile resintering Both FP gas release and swelling were small in Zone IV.73

Temperature range of each zone found in (U,Pu)N fuel irradiated in DFR to 4% FIMA was roughly evaluated by Matzke1as follows Zone I appeared at

a temperature higher than 1673 K, while Zone IV predominated at a temperature lower than 1423 K The intermediate temperature ranging from 400 to

500 K corresponded to Zone III On the other hand, under the ‘cold fuel concept’ proposed by Blank,69 most part of fuel pellets should represent the as-fabricated structure seen in Zone IV characterized

by low FP gas release and mild swelling

Richter et al.74

observed the macro- and microstructures of unirradiated (U,Pu)N pellets

Zone

Low temperatures, edge

High temperatures, center

Structure of as-fabricated fuel

Grain growth, grain boundary bubbles

Pseudocolumnar grain zone, elongated grains and pores

Very porous central zone

Figure 5 Schematic presentation of structural zones observed in MX-type fuels Reproduced from Matzke, Hj Science of Advanced LMFBR Fuels; North-Holland: Amsterdam, 1986.

heated in He or N2atmosphere under temperature

gradient From the viewpoint of structural stability,

they suggested that the operational limit of

tempera-ture for (U,Pu)N was 2000 K in normal condition,

although the structural change observed was affected

by the oxygen impurity contents

During the power rise at an early stage of

irradia-tion, tensile stresses are created in the outer zone and

compressive stresses at the central zone of fuel

pel-lets Since most of ceramics are more sensitive to

tensile than compressive stresses, crack formation

usually occurs at the periphery during start up It

was reported that radial cracks predominated in the

Na-bonded (U,Pu)N fuel pellets irradiated in

ther-mal test reactors, while circumferential cracks

pre-dominated in the He-bonded (U,Pu)N fuel pellets.62

On the other hand, many short radial cracks were

observed at the periphery in the He-bonded (U,Pu)N

fuel pellets irradiated in EBR-II.58Healing of cracks

was often observed in the He-bonded (U,Pu)N fuel

after the closure of He gap by fuel swelling

3.02.3.5 FP Gas Release

Since the number of nitride fuel pins subjected to PIE

is limited, there have been no systematic results

dealing with FP gas release of nitride fuel But it is

generally known that FP gas release of nitride fuel is

much lower than that of MOX fuel The FP gas

release will be influenced by burnup, pellet density,

grain size, and the characteristics of porosities as well

as fuel temperature By statistically dealing with the

data reported for 95 UN and 39 (U,Pu)N fuels

Storms10proposed an equation for FP gas release of

nitride fuel as a function of fuel temperature, burnup,

and density as follows:

R ¼ 100= exp 0:0025ð90D  0 :77=Bu0 :09 TÞþ 1

½3

where R is the FP gas release rate (%), D the fuel

pellets density (% TD), Bu the burnup (% FIMA),

andT the temperature of fuel (K)

Bauer et al.62

summarized the results of FP gas release for (U,Pu)N fuel irradiated at a relatively

low fuel temperature From the fuel porosity

depen-dence, they suggested that the recoil of FP gas atom

from the geometric surface was responsible for the

release from pellets with a density higher than 85%

TD, and it rapidly increased with the decrease of

density lower than 82% TD since the release through

the surface connected porosity became responsible

Figure 6 shows the results for (U,Pu)N fuel

irradiated in JOYO to 4.3% FIMA75 in comparison with the porosity dependence reported by Bauer

et al.62 Furthermore, for the (U,Pu)N fuel pellets irradiated in JOYO, Tanaka et al suggested that about 80% of FP gas was still retained in the intra-granular region, about 15% was in the gas bubbles, and about 5% was released from the pellets, based on the results of pin puncture test and electron probe microanalysis (EPMA) of fuel pellets

On the other hand, Coquerelle et al.72

reported that the release rate of Xe from the central region of (U,Pu)N pellets was about 45% and about 15% from the outer part of the fuel pellets irradiated in DFR Although the burnup was low in their irradiation campaign, FP gas release was relatively high because

of a linear power higher than 100 kW m1 Therefore, the high FP gas release could be explained by the diffusion process as in the MOX fuel

3.02.3.6 Swelling and FCMI

As mentioned above, FP gas release of nitride fuel is low in general This characteristic potentially leads to

0.1 1 10 100

L414

L413

Estimated recoil release from geometric surface

40 30

20 10

0

Fuel porosity (%)

Release through surface-connected porosity

This study

Figure 6 Porosity dependence of FP gas release rate of (U,Pu)N fuel Reproduced from Tanaka, K.; Maeda, K.; Katsuyama, K.; et al J Nucl Mater 2004, 327, 77–87.